On Fri, Sep 05, 2003 at 11:25:23AM -0700, Ray Saintonge wrote:
Choosing the top ten for the short list was based on an arbitrary number. If 10th and 11th place are close in votes why not include the 11th? If the occasional bit of improper voting happens, it is then marginalized. It would take massive vote fraud to make a difference. I prefer marginalizing the casual cheater to spending enormous energy developing elaborate systems to catch them.
As for what to do with the short list (which also includes candidates with multiple variants) perhaps they should be first reviewed for technical workability. Some of the high vote getters strike me as very complicated designs that may not scale down very well for low-resolution situations.
*IF* we use Condorcet method, then you're right - one more option doesn't change anything. But if we go with some voting method prone to tactical voting, like FPTP, Average, or IRV, then it isn't so simple.